360Brew explained: LinkedIn's new algorithm, in plain English

In late 2025, LinkedIn quietly switched off most of its old algorithm. The system that used to count likes and hashtags is gone. In its place sits 360Brew, a 150-billion-parameter AI model that reads what you actually write, who you really are, and whether anyone genuinely cares. If your reach has dropped, your engagement has fallen off, or your posts feel like they're going nowhere, that's probably why.

360Brew is LinkedIn's new brain. It reads context, not engagement.

For years, LinkedIn used a simple ranking system. It counted likes. It counted hashtags. It pushed posts that got fast engagement.

360Brew is the replacement. It's a single AI model, similar in scale to ChatGPT but trained on LinkedIn data, running across the entire platform. Your feed, your job suggestions, even your ads. All of it now flows through one brain.

What makes it different is what it actually does. The old algorithm counted reactions. 360Brew reads what you wrote. It checks your profile. It looks at whether the people engaging with your posts are credible in the field you're posting about. It decides if your post helped someone, or just wasted their time.

The signals that worked before mostly don't anymore. Hashtag stuffing is penalised. Engagement pods get shadow-banned. Posting five generic things a week is beaten by one substantial post.

The good news is that the rules are simpler than they look. Write things that matter, in your actual area of expertise, for an audience that genuinely cares. That's it. The rest of this article is just the detail of how that plays out.

The old algorithm counted reactions. The new one reads context.

Before 360Brew, the algorithm worked like a calculator. Likes added points. Hashtags added points. Fast engagement added more points. Whatever scored highest got pushed to more feeds. Easy to game, which is exactly why it eventually broke.

360Brew works differently. It reads the actual post the same way a person would. It understands what you're talking about, decides whether the topic matches your stated expertise, and judges whether the people engaging with you are credible in that area.

That's a fundamental change. There's nothing to optimise that isn't, fundamentally, just writing well about something you actually know.

It judges you on four things, all the time.

Researchers who've studied how 360Brew ranks posts have identified four things it consistently weighs:

1. Profile coherence. Does your post line up with what your profile says you do? Headline, About section, recent activity. They all need to tell the same story.
2. Network relevance. Are the people interacting with your posts actually credible in your field? A post about HR strategy that gets liked by 50 people in unrelated industries scores worse than one that gets 5 thoughtful comments from senior HR leaders.
3. Engagement quality. Saves matter more than likes. Comments with substance matter more than emoji reactions. Time spent reading matters more than scroll-pasts.
4. Topic consistency. 360Brew watches what you post about over time, and evidence suggests it can take up to 90 days for the algorithm to decide what your professional territory actually is. Keep changing topics, and you keep restarting that clock.

What 360Brew rewards (and how much).

The numbers below come from analyses by Richard van der Blom (1.8 million posts), AuthoredUp (over 3 million posts) and Botdog. The figures vary slightly between sources, but they all point the same way.

• Native formats win. Document posts and carousels both pull around 6.6% engagement. Vertical video under 60 seconds retains 87% of viewers.
• External links cost reach. Estimates from 25% to 60%. All of it bad.
• Credible peers count more. Five thoughtful comments from real industry people carry roughly 5× the weight of fifty reactions from random connections.

And what kills your reach (sometimes permanently).

The list of things that used to work and now actively hurt you:

• Hashtag stuffing. Some analyses say 3+ hashtags drops reach by 70%; others put the spam-filter trigger at 5+. Direction is the same: use fewer, or none. LinkedIn has actively removed hashtag-following from the platform.
• Engagement pods. LinkedIn's product team has openly stated their goal is to make pods "entirely ineffective". Detection is automated. Penalties run from 60–90 day shadow bans up to permanent suspension.
• External links in the body of posts. Reach penalty of 25% to 60%. Putting the link in the first comment is also detected now.
• Multiple posts within 24 hours. Posting twice in a single day actively suppresses the newer post. Space your posts at least 24 hours apart.
• Off-topic posts. A "SaaS expert" who posts a motivational quote about Monday mornings will get suppressed. The post doesn't match what the algorithm thinks you do for a living.
• Generic AI templates. It isn't AI itself that's penalised. It's templated, formulaic structures the algorithm can pattern-match.
• Same format twice in a row. Two carousels back-to-back can suppress reach by around 20%. Mix your formats.

Most LinkedIn advice is still optimising for an algorithm that doesn't exist.

Most articles, courses, and LinkedIn coaches haven't caught up. They're still teaching tactics that worked for the algorithm LinkedIn ran in 2023.

• "Use 5 hashtags." Already covered. Use fewer, or none.
• "Engage with 50 people a day, especially in pods." The first half is fine if you do it genuinely. The pod part is detected and punished.
• "Use this exact 7-line hook formula." Hooks aren't wrong. Formulas are. The algorithm pattern-matches templated openings.
• "Post motivational content for engagement." Off-topic posts hurt your topic authority, even if they get reactions.
• "Post 5 times a week." Volume isn't the problem; substance, spacing, and topic consistency are. 2–3 substantive on-topic posts a week is where most data converges.

How ContentLark is built around the rules that actually apply.

ContentLark exists because the rules of LinkedIn changed and most tools didn't. Every part of how it works maps directly onto what 360Brew rewards or penalises.

• The audit measures all four pillars. Profile coherence, network relevance, engagement quality, topic consistency. You get a single Authority Score and a line-by-line breakdown.
• The strategy layer builds your topic territory. 2–3 content pillars to stick to. Same pillars, every post. That's how the 60–90 day topic authority signal builds.
• The content engine writes for the new signals. Substance and saves. Native formats only. No external links in the body. No hashtag spam. Smart scheduling spaces posts properly.
• The audience engine grows credible network relevance. Connection campaigns target your ICP and engage warmly with people who already interact with your content.
• The analytics dashboard tracks your Authority Score over time. Not just likes. The signals that matter, week by week.